Reference-free lossless compression of nanopore sequencing reads using an approximate assembly approach

Area of Science:

• Genomics
• Bioinformatics
• Data Compression

Background:

• Genome sequencing data volume is rapidly increasing, necessitating efficient data compression.
• Nanopore sequencing offers portability and long reads, but specialized compression tools are lacking.
• Existing general-purpose or short-read compressors are suboptimal for nanopore FASTQ data.

Purpose of the Study:

• Introduce NanoSpring, a novel reference-free compressor for nanopore sequencing reads.
• Evaluate NanoSpring's compression performance and speed against existing methods.
• Provide an efficient solution for storing, transferring, and analyzing large nanopore datasets.

Main Methods:

• Developed NanoSpring utilizing an approximate assembly approach.
• Tested NanoSpring on diverse datasets: bacterial, metagenomic, plant, animal, and human whole genomes.
• Focused compression on base sequences within FASTQ files.

Main Results:

• Achieved 0.35-0.65 bits per base compression for high-quality nanopore data, outperforming gzip by 3-6x.
• Demonstrated competitive compression ratio and resource usage compared to CoLoRd.
• Showcased significantly faster decompression (over 4x with 20 threads) than CoLoRd.

Conclusions:

• NanoSpring provides superior compression efficiency for nanopore sequencing data.
• The tool offers a significant speed advantage in decompression, crucial for real-time analysis.
• NanoSpring is a valuable asset for managing and analyzing the growing volume of nanopore sequencing data.